DE69703546T2 - FIXED BED OF NANO FIBERS WITH IMPROVED FLOW CHARACTERISTICS - Google Patents

Abstract

The general area of this invention relates to porous materials made from nanofiber packed beds. More particularly, the invention relates to altering the porosity or packing structure of a nanofiber packed bed structure by blending nanofibers with scaffold particulates having larger dimensions. For example, adding large diameter fibers to a nanotube packed bed to serve as a scaffolding to hold the smaller nanofibers apart and prevent the nanofiber bed structure from collapsing. This increases the average pore size of the mass by changing the pore size distribution and alters the packing structure of the packed bed. The increase in average pore size is caused by the creation of larger channels which improves the flow of liquids or gasses through these materials.

Description

Field of the invention

The invention relates generally to nanofiber packed beds with improved liquid flow properties and to methods of making and using the same. More particularly, the invention relates to nanofibers that are uniformly or non-uniformly mixed with supporting scaffold particles to form packed beds that have increased liquid flow rates and increased overall average pore size. More particularly, the invention relates to the use of such packed beds for a variety of uses including products such as flow electrodes, chromatography media, adsorbents, and filters.

Description of the state of the art

Interwoven nanofiber mats and nanofiber systems or compositions have previously been prepared to take advantage of the increased surface area per gram achieved by using extremely thin fiber diameters. These earlier mats or compositions are either in the form of compact, dense mats of interwoven fibers and/or are limited to microscopic structures (i.e., with a largest dimension of less than 1 micron). Interwoven nanofiber mats or nanofiber compositions or systems have previously been prepared by dispersing nanofibers in aqueous or organic media, followed by filtering the nanofibers to form a mat. The mats have also been prepared by preparing a gel or paste of carbon fibrils in a fluid, e.g., an organic solvent such as propane, and then heating the gel or paste to a temperature above the critical temperature of the medium, removing the supercritical fluid, and finally removing the resulting porous mat or plug from the vessel in which the process was carried out. See U.S. Patent Application No. 08/428,496, entitled "Three Dimensional Macroscopic Assemblages of Randomly Oriented Carbon Fibrils and Composites Containing Same" by Tennent et al., which is hereby incorporated by reference.

A disadvantage of the previous compositions or mats made by the methods described above is poor fluid permeability within the structure. When the nanofiber suspensions are dried from the suspension liquid, particularly water, the surface tension of the liquid tends to pull the nanofibers together into a tightly packed "mat." Alternatively, the structure may simply collapse. The pore size of the resulting mat is determined by the spacing between the fibers, which tend to be quite small due to the compression of these mats. As a result, the fluid permeability of such mats is poor.

Therefore, although previous work has shown that nanofibers can be assembled into packed, thin, membrane-like compositions or systems through which a fluid can flow, the small diameters of the nanofibers result in a very small pore structure that presents a large resistance to fluid flow.

It would be desirable to overcome the above disadvantages by producing a porous packed bed with improved fluid flow and an altered pore size distribution, since there are applications for porous nanofiber packed beds that require fluid flow and the resistance to fluid transport creates serious limitations and/or disadvantages for such applications. The improved fluid flow properties provided by these invention make such applications more possible and/or more efficient.

TASK OF THE INVENTION

It is therefore an object of this invention to provide porous nanofiber packed bed structures with improved fluid flow properties and/or increased average pore size.

It is a further object of the invention to provide a material composition comprising a three-dimensional, macroscopic nanofiber filling bed consisting of a mixture of randomly oriented nanofibers and larger scaffold particles.

It is a further object of the invention to provide methods for making nanofiber packed beds with improved liquid flow properties and methods of using them. It is also a further object of the invention to provide improved filter media, chromatography media or supports, adsorbents, electrodes, EMI shields and other compositions of industrial value based on three-dimensional porous nanofiber packed beds.

The above and other objects and advantages of the invention are disclosed in or will be apparent from the following description and drawings.

SUMMARY OF THE INVENTION

The general scope of this invention relates to porous materials made from nanofiber infill beds. In particular, the invention relates to altering the porosity or the packing structure of a nanofiber packed bed structure by mixing nanofibers with scaffold particles that have larger dimensions. For example, adding larger diameter fibers to a nanotube packed bed acts as a scaffold that keeps the smaller nanofibers spaced apart and prevents the nanofiber packed bed structure from collapsing. This increases the average pore size of the mat by changing the pore size distribution and changes the packing structure of the packed bed. The increase in average pore size is caused by the creation of larger channels, which improves the flow of liquids or gases through these materials.

Accordingly, it is a concern of the invention to alter the average pore size and packing structure of packed nanofiber layers by admixing larger particles, preferably fibers with larger diameters. The larger particles alter the packing of the nanofibers and result in structures with reduced resistance to fluid permeation. The present invention provides the unexpected advantage of being able to form a packed bed structure of nanofibers with improved fluid permeation properties as a result of the framework effect caused by the framework particles.

BRIEF DESCRIPTION OF THE DRAWINGS

Figure 1 is a light micrograph (magnification x 50,000) of a nanofiber packing bed showing a region of a nanofiber mat comprising randomly oriented, interwoven carbon fibrils.

Figure 2 is a light micrograph (magnification x 2,000) showing a packing bed comprising scaffold fiber particles and web-like structures of randomly oriented, interwoven carbon fibrils.

Fig. 3 is a light micrograph (magnification x 200) showing scaffold particles in the form of fibers (4-8 µm diameter) and tissue-like regions of fibril mats.

Figure 4 is an optical micrograph (· 100) of a packed bed structure according to the present invention.

Figure 5 shows a graphical representation of the flow rate/mat structure relationship for nanofiber mats (without scaffold) as a comparison, where the vertical axis represents the flow rate and the horizontal axis represents the flow of the total volume through the mat for mats with different thickness and density.

Figure 6 shows a graphical representation of the flow rate/mat density relationship for carbon nanofiber mats (without scaffold) for comparison, where the vertical axis represents the flow rate normalized to the mat thickness and the horizontal axis represents the mat density.

Fig. 7 shows a graphical representation of the flow rate / nanofiber fraction relationship of an embodiment of the invention, wherein the vertical axis represents the flow rate and the horizontal axis represents the weight fraction of nanofibers in a packed bed containing carbon fibrils and carbon fibers.

Figure 8 is a graphical representation of the current-voltage relationship for the electrode of Example 8, where the vertical axis represents the current and the horizontal axis represents the applied potential at several potential scanning speeds.

DETAILED DESCRIPTION OF THE INVENTION Definitions

The term "fluid flow rate property" refers to the ability of a liquid or gas to flow through a solid structure. For example, the rate at which a volume of liquid or gas flows through a three-dimensional structure with a specific cross-sectional area and specific thickness or height (i.e., milliliters per minute per square centimeter per mil of thickness) at a fixed pressure differential across the structure.

The term "isotropic" means that all measurements of a physical property within a plane or volume of the packed bed, regardless of the direction of measurement, give a constant value. It is understood that measurements of such non-solid compositions must be made on a representative packed bed sample so that the average value of the voids is taken into account.

The term "macroscopic" refers to structures that have at least two dimensions larger than 1 mm.

The term "nanofiber" refers to elongated structures with a cross-section (angular fibers with edges) or diameter (rounded) of less than 1 micron. The structure can be either hollow or solid. This term is defined below.

The terms "packed bed", "composition" or "mat" refer to a structure comprising an arrangement of a mat of interwoven individual nanofibers, scaffold fibers and/or scaffold particles. The term "packed bed" is hereinafter understood to include and be interchangeable with the terms "mats", "composites" and related three-dimensional structures. The term "packed bed" does not include loose masses of particles.

The term "packing structure" refers to the internal structure of a packed bed, including the relative orientation of the fibers, the versatility and overall average of the fiber orientation, the proximity of the fibers to each other, the voids or pores created by the gaps and spaces between the fibers, and the size, shape, number and orientation of the flow channels or paths formed by the interconnection of the voids or pores. The term "relative orientation" refers to the orientation of an individual fiber with respect to the other fibers (i.e., aligned versus non-aligned). The "versatility" and "overall average" of the fiber orientations refer to the range of fiber orientations in the packed bed (alignment and orientation with respect to the external surface of the bed).

The term "physical property" means an inherent, measurable property of the porous packing bed, e.g. resistance, fluid flow properties, density, porosity, etc.

The term "relative" means that 95% of the values of the physical property lie within plus or minus 50% of a mean value when measured along an axis or within a plane or volume of the structure.

The term "scaffold particles" refers to solid particles that are capable of causing a scaffold effect when mixed with nanofibers. At least one dimension of the "scaffold particle" is substantially larger than at least one dimension of the nanofibers. The "scaffold particles" can have various three-dimensional shapes including fibers, cubes, platelets, disks, etc. "Scaffold particles" are discussed further below.

The term "substantially" means that 95% of the values of the physical property are within plus or minus 10% of a mean value when measured along an axis or within a plane or volume of the structure.

The terms "substantially isotropic" or "relatively isotropic" correspond to the ranges of variation of the values of a physical property as specified above.

Nanofibers

The term nanofibers refers to various fibers with very small diameters including fibrils, whiskers, nanotubes, buckytubes, etc. Due to their size, such structures provide significant surface area when inserted into a packed bed structure. In addition, such a structure can be manufactured with high purity and uniformity. Preferably, the nanofiber used in the present invention has a diameter of less than about 1 µm, preferably less than about 0.5 µm, and more preferably less than 0.1 µm, and most preferably less than 0.05 µm.

The fibrils, buckytubes, nanotubes and whiskers referred to in this application are distinguishable from continuous carbon fibers which are commercially available as reinforcement materials. Unlike nanofibers, which have desirably large but inevitably finite length-to-width ratios, continuous carbon fibers have length-to-width ratios (L/D) of at least 104 and often 106 or more. The diameter of continuous fibers is also far larger than that of fibrils and is always greater than 1.0 µm and is typically 5 to 7 µm.

Continuous carbon fibers are produced by the pyrolysis of organic precursor fibers, usually rayon, polyacrylonitrile (PAN), and pitch. Consequently, they can contain heteroatoms in their structure. The graphitic nature of "virgin" continuous carbon fibers varies, but they can be subjected to a subsequent graphitization step. Differences in the extent of graphitization, orientation and crystallinity of the graphite planes, if present, the potential presence of heteroatoms, and even the absolute difference in substrate diameter mean that experience with continuous fibers is poorly predictive of nanofiber chemistry. The various types of nanofibers suitable for use in porous packed bed structures are discussed below.

Carbon fibrils are worm-like carbon deposits with diameters of less than 1.0 µm, preferably less than 0.5 µm, more preferably less than 0.2 µm, and most preferably less than 0.05 µm. They exist in a variety of forms and are produced by catalytic decomposition of various carbon-containing gases on metal surfaces. Such worm-like carbon deposits have been observed since the beginning of electron microscopy. A good early review and reference can be found in Baker and Harris, Chemistry and Physics of Carbon, eds. Walker and Thrower, vol. 14, 1978, p. 83, and N. Rodriguez, J. Mater. Research, Vol. 8, p. 3233 (1993), which are hereby incorporated by reference (see also A. Obelin and M. Endo, J. of Crystal Growth, Vol. 32 (1976), pp. 335-349, hereby incorporated by reference).

U.S. Patent No. 4,663,230 to Tennent, hereby incorporated by reference, describes carbon fibrils that are free of a continuous thermal carbon coating and that have multiple ordered graphitic outer layers that are substantially parallel to the fibril axis. Thus, they can be characterized such that their c-axes, the axes that are perpendicular to the tangents of the curved graphite layers, are substantially perpendicular to their inner They generally have diameters of no more than 0.1 µm and length-to-diameter ratios of at least 5. Desirably, they are substantially free of a continuous thermal carbon coating, i.e., pyrolytically deposited carbon resulting from thermal cracking of the gas stream used to make them. The Tennent invention provides access to fibrils of smaller diameter, typically 35 to 700 Å (0.0035 to 0.070 µm), and to an ordered, "as grown" graphitic surface. Fibrillar carbons of less perfect structure, but also without a pyrolytic outer carbon layer, have also been grown.

U.S. Patent No. 5,171,560 to Tennent et al., hereby incorporated by reference, describes carbon fibrils without thermal coating and having graphitic layers that are substantially parallel to the fibril axis such that the projection of these layers onto the fibril axes extends over a distance of at least two fibril diameters. Typically, such fibrils are substantially cylindrical graphitic nanotubes with a substantially constant diameter and comprise cylindrical graphitic layers whose c-axes are substantially perpendicular to their cylindrical axes. They are substantially free of pyrolytically deposited carbon, have a diameter of less than 0.1 µm and a length-to-diameter ratio of greater than 5. These fibrils are of primary interest in the invention.

Further details regarding the formation of carbon fibril aggregates can be found in the disclosure of U.S. Patent 5,165,909 to Tennent; Snyder et al., U.S. Patent Application No. 149,573, filed January 28, 1988; and PCT Application No. US89/00322, filed January 28, 1989 ("Carbon Fibrils") WO 89/07163; and Moy et al., U.S. Patent Application No. 413,837, filed September 28, 1989; and PCT Application No. US90/05498, filed September 27, 1990 ("Fibril Aggregates and Method of Making Same") WO 91/05089; and U.S. No. 08/479,864 by Mandeville et al., filed June 7, 1995; and No. 08/329,774 by Bening et al., filed October 27, 1984; and No. 08/284,917, filed August 2, 1994; and No. 07/320,564, filed October 11, 1994 by Moy et al., all of which, as with the present invention, are assigned to the same assignee and are hereby incorporated by reference.

Moy et al., U.S. Application No. 07/887,307, filed May 22, 1992, hereby incorporated by reference, describes fibrils prepared as aggregates with various morphologies (as determined by scanning electron microscopy) where they are randomly intertwined to form intertwined fibril balls resembling bird's nests (BN); or prepared as aggregates consisting of bundles of straight or slightly bent or kinked carbon fibrils with substantially the same relative orientation and having the appearance of combed yarn (cy), e.g., the longitudinal axis of each fibril (despite individual bends or kinks) extends in the same direction as the surrounding fibrils in the bundles; or be prepared as aggregates consisting of straight to slightly curved or kinked fibrils loosely interwoven to form an open net (ON) structure. In open net structures, a greater degree of fibril intertwining is observed than in combed yarn aggregates (in which the individual fibrils have essentially the same relative orientation), but less than in the case of bird's nests. Combed yarn aggregates and open net aggregates are more easily dispersed than bird's nest aggregates, making them useful in the preparation of composites where uniform properties throughout the structure are desired.

If the projection of the graphite-like layers onto the fibril axis extends over a distance of less than two fibril diameters, the carbon layers of the graphite-like nanofiber take on an appearance in the form of of fish bones. These are referred to as fish bone fibrils. Geus, U.S. Patent No. 4,855,091, hereby incorporated by reference, provides a process for the preparation of fish bone fibrils that are substantially free of a pyrolytic coating. These fibrils are also suitable for use in the invention.

McCarthy et al., U.S. Patent Application No. 351,967, filed May 15, 1989, hereby incorporated by reference, describes methods for oxidizing the surface of carbon fibrils which involve contacting the fibrils with an oxidizing agent comprising sulfuric acid (H2SO4) and potassium chlorate (KClO3) under reaction conditions (e.g., time, temperature, and pressure) sufficient to oxidize the fibril surface. The fibrils oxidized according to the methods of McCarthy et al. are non-uniformly oxidized, that is, the carbon atoms are substituted with a mixture of carboxyl, aldehyde, ketone, phenol, and other carbonyl groups.

Fibrils were also non-uniformly oxidized by treatment with nitric acid. The international application PCT/US94/10168 discloses the formation of oxidized fibrils containing a mixture of functional groups.

In a published work, McCarthy and Bening (Polymer Preprints ACS Div. of Polymer Chem. 30 (1) 420 (1990)) prepared derivatives of oxidized fibrils to show that the surface contained a variety of oxidized groups. The compounds they prepared, phenylhydrazones, haloaromatic esters, thallium salts, etc., were chosen for their analytical utility, for example, because they were strongly colored or had other strong and easily identified and distinctive signals. These compounds were not isolated and, unlike the derivatives described here, are of no practical importance.

Carbon nanotubes with a morphology similar to that of the catalytically grown fibrils described above were grown in a high temperature carbon arc (Iijima, Nature, 354, 56, 1991, hereby incorporated by reference). It is now generally accepted (Weaver, Science, 265, 1994, hereby incorporated by reference) that these arc-grown nanofibers have the same morphology as the earlier catalytically grown fibrils of Tennent. Arc-grown carbon nanofibers are also useful in the invention.

Framework particles

Scaffold particles are particulate solids with a shape and size suitable to provide a scaffold effect when mixed with nanofibers. The scaffold particles are of a shape and size such that they disrupt the packing structure of the nanofibers. This results in a packed bed with an increased average pore size. The scaffold effect increases the number of large pores and the average pore size, which in turn increases the bed throughput. The scaffold particles are used as a diluent and/or as a mechanically stronger scaffold, helping to overcome the forces of surface tension during the drying process, reducing the density of the nanofiber fraction of the resulting mat as a composite.

Preferably, the scaffold particles have at least one dimension that is larger than the largest dimension of the nanofibers and/or at least one second largest dimension that is larger than the second largest dimension of the nanofiber. The largest dimension of the scaffold particle may be comparable to the largest dimension of the nanofiber. For example, nanofibers and fatty fibers of the same length may be used, as long as the diameters of the fatty fibers are substantially larger than the diameters of the nanofibers.

Preferably, the largest dimension of the scaffold particle is at least 10 times larger than the largest dimension of the nanofibers, more preferably 50 times larger, more preferably 100 times larger, and most preferably 200 times larger.

Preferably, the second largest dimension of the scaffold particle is at least 10 times larger than the second largest dimension of the nanofibers, more preferably 50 times larger, more preferably 100 times larger, and most preferably 200 times larger.

The scaffold particles also preferably have a largest dimension (e.g., length of a fiber) of greater than 1 µm, more preferably greater than 5 µm, even more preferably greater than 10 µm, and most preferably greater than 50 µm. The scaffold particles preferably have a second largest dimension (e.g., diameter of a fiber or thickness of a disk) of greater than 0.1 µm, more preferably greater than 1 µm, even more preferably greater than 3 µm, and most preferably greater than 5 µm.

The shape of the scaffold particles can take many forms as long as the particle provides a sufficient scaffold effect to improve the fluid flow properties by the desired amount. Suitable shapes include fibers, platelets, disks, cones, pyramids, cubes, irregularly shaped solid particles, etc.

Preferably, the scaffold particles are in the form of a fiber. The diameter of the scaffold particle fiber is preferably at least 10 times larger than the nanofiber diameter, more preferably 50 times larger, even more preferably 100 times larger, and most preferably 200 times larger.

The surface chemistry, structure or composition of the scaffold particle can also be modified to improve or adjust the scaffold effect. For example, particles with a rough surface can cause an increasing scaffold effect, since such particles would be better able to keep the nanofibers away from each other. According to one embodiment of the invention, more than one type of scaffold particle is introduced into the packed bed.

The framework particles can be, for example, polymeric, inorganic, glassy or metallic. The particle can have the same or a different composition with respect to the nanofiber. Preferably, the framework particles are either glass fiber particles or carbon fibers.

According to a preferred embodiment, carbon fibers are used as scaffold particles. The carbon fibers have the advantage for this purpose of being conductive. Metal fibers with even greater conductivity can also be mixed in for the same reason. Consequently, carbon fibers, which are conductive but also consist of carbon, can be used as scaffold particles and mixed with carbon nanofibers to create a carbon-based packed bed that has improved conductivity, improved fluid permeability and high carbon purity.

Nanofiber filling beds with improved fluid flow

The general scope of this invention relates to porous packed bed materials made from packed nanofibers. The invention relates to a method for altering the porosity of a packed bed structure by mixing nanofibers with other larger diameter fibers or particulate materials that serve as a scaffold that holds the smaller nanofibers apart and prevents them from collapsing. The creation of larger channels within these composite materials improves the flow of liquids or gases through these materials.

The flow of a fluid through a capillary is described by the Poiseuille equation, which relates the flow rate to the pressure difference, the fluid viscosity, the path length and the size of the capillaries. The flow rate per unit area varies with the square of the pore size. Accordingly, doubling the pore size leads to a quadrupling of the flow rates. The creation of much larger pores in the nanofiber packed bed structure leads to increased liquid flow, since the flow through the larger pores is much greater.

As stated above, nanofiber packed beds that are not provided with a scaffolding mechanism result in poor fluid flow properties. This is especially true when liquid nanofiber suspensions are dried from the suspending fluid and the surface tension of the liquid tends to push the nanofibers together into a tightly packed "mat" where the pore size, determined by the space between the fibers, tends to be quite small. This results in a packed bed with poor fluid flow properties.

When the nanofiber packing beds are formed without the scaffold particles, they collapse to form dense mats. The scaffold particles hold the mats upright and break or tear these mats apart to form discontinuous nanofiber regions and void channels between these regions. Referring to Fig. 1, the micrograph shows the dense packing of nanofibers in the nanofiber region or nanofiber domain of the mat. Fig. 2 is a lower magnification micrograph showing the nanofiber domains adhering to the scaffold particles and forming "web-like" structures between and over them. The interwoven fibrous structures of the nanofibers cause them to adhere to one another to form the web-like or felt-like structure rather than completely breaking apart into individual particles. If non-fibrous nanoparticles (e.g. spheres) were used, these particles would simply fall through the structure and settle at the bottom. Figures 3 and 4 are low-magnification micrographs of packings containing carbon fibrils and carbon fibers.

Surprisingly, packing beds comprising larger scaffold particles mixed with nanofibers show improved flow properties. This allows the high surface area of the nanofibers to be readily exploited in situations where volume transport of a fluid or gas through the material is required.

The improved packing structure of the packed bed provides large flow channels, which allows a large surface area of the nanofibers to be accessible. This means that the nanofibers forming the outer walls of or in contact with the large flow channels formed in the composite structure allow an increase in the accessible nanofiber surface area.

One aspect of the invention relates to a material composition comprising a porous packed bed with a plurality of nanofibers and a number of scaffold particles mixed to form a porous structure. Preferably, the packed bed has an improved liquid flow characteristic. Preferably, the liquid flow rate of water is greater than 0.5 ml/minute/cm2 at a pressure differential across the packed bed of about 1 atm for a bed having a thickness of 1 mil, more preferably greater than 1.0 ml/minute/cm2.

Generally, the invention is a composition of matter consisting essentially of a three-dimensional macroscopic system of a plurality of randomly oriented nanofibers mixed with scaffold particles. Preferably, the beds have at least two dimensions greater than 1 mm, more preferably greater than 5 mm. Preferably, the resulting packed beds have a bulk density of 0.1 to 0.5 gm/cc.

The packing bed made according to the invention has structural and mechanical integrity. This means that the beds can be applied without breaking or falling apart, although the beds may have some flexibility. Due to the fact that the nanofibers can be packed to form a random, Because the particles are interwoven into a fabric-like structure, the filling beds have a felt-like character. Compared to simply loose particle masses, the filling beds have significantly higher strength and resistance.

Preferably, the packed bed according to the invention has at least one fluid flow property that is greater than that of a nanofiber packed bed without scaffold particles. This means that the addition of the scaffold particles improves the fluid flow rate, for example, compared to a nanofiber packed bed without scaffold particles. Preferably, the scaffold addition increases the flow rate by a factor of at least 2, more preferably by a factor of at least 5, even more preferably by a factor of at least 10, and most preferably by a factor of at least 50.

According to a preferred embodiment, the packing bed has a surface area of more than about 25 m²/g.

Preferably, the packed bed has relatively or substantially uniform physical properties along at least one dimensional or spatial axis and desirably has relatively or substantially uniform physical properties in one or more planes within the packed bed, i.e., has isotropic physical properties in that plane. In other embodiments, the entire packed bed is relatively or substantially isotropic with respect to one or more of its physical properties.

Preferably, the packed bed has substantially isotropic physical properties in at least two dimensions. When fibers or platelets are used as scaffold particles, they tend to align in a single plane. However, the resulting structure is isotropic in the plane of alignment. The packed bed can have a uniform or non-uniform distribution of the scaffold particles and nanofibers.

According to one embodiment of the invention, the distribution of the scaffold particles in the nanofiber packed bed is non-uniform. As can be seen from comparing Figures 1, 2, 3 and 4, high magnifications show that the nanofibers can cluster and form web-like regions, although the distribution may appear uniform at lower magnification. Although this can occur, the macroscopic properties of the material may be relatively uniform, as shown in Figure 4. Alternatively, the packing structure may be non-uniform. For example, a thin layer region with a higher concentration of nanotubes may be formed on the upper region of the packed bed. Alternatively, the thin layer of nanofibers may be formed in the lower region or within the packed bed structure. The distribution of the scaffold particles may be varied to change the properties of the packed bed.

The average pore size and overall packing structure of the packed bed can be tuned by varying several parameters. These parameters include: (a) weight percent of nanofiber and/or support particles, (b) size, shape and surface properties of the nanofiber and/or scaffold particle, (c) the composition of the nanofiber and/or scaffold particle (e.g., carbon versus metal), and (d) the method of preparing the packed bed.

Although the spaces between the nanofibers are irregular in size and shape, they can be considered as pores and characterized by methods used for characterizing porous media. The size distribution of the interstices and voids in such networks can be controlled by the concentration and extent of dispersion of nanofibers mixed with the scaffold. Scaffold addition to the nanofiber packings causes a change in the pore size distribution. Although the pore size in the nanofiber mat domains is not significantly changed, breaking away these regions results in larger void channels between the nanofiber regions. The result is a bimodal pore size distribution: small clean spaces with the nanofiber areas and a large void between these areas.

According to one embodiment, the porous packing beds may contain an amount of nanofibers in the range of 1 wt% to 99 wt%, preferably 1 wt% to 95 wt%, more preferably 1 wt% to 50 wt%, and most preferably 5 wt% to 50 wt%. The corresponding amount of scaffold particles is in the range of 99 to 1 wt%, more preferably 99 to 5 wt%, even more preferably 99 to 50 wt%, and most preferably 95 to 50 wt%.

According to another embodiment, the porous packing beds may comprise a solid non-void volume with an amount of nanofibers in the range of 1 vol.% to 99 vol.%, preferably 1 vol.% to 95 vol.%, more preferably 1 vol.% to 50 vol.%, and most preferably 5 vol.% to 50 vol.%. The corresponding amount of scaffold particles is in the range of 99 to 1 vol.% (of solid non-void volume), more preferably 99 to 5 vol.%, more preferably 99 to 50 vol.%, and most preferably 95 to 50 vol.%.

As discussed above, varying the amount of nanofibers and scaffold particles changes the packing structure and consequently changes the fluid flow properties. Figure 5 is a graphical representation of the relationship between water fluid flow and mat thickness for a carbon fibril mat (2 cm2) without scaffold particles. Figure 6 is a graphical representation of flow rate versus mat density. As can be seen from the graph, the fluid flow properties for these mats are poor. Flow rate decreases exponentially with increasing mat density, corresponding to a decrease in pore volume. Figure 7 is a graphical representation of the relationship of fluid flow rate/nanofiber fraction for a packed bed (2 cm2) made according to the present invention. As can be seen from the graph can be removed, increased throughput speeds are achieved by adding scaffold fibers to a nanofiber mat.

According to a preferred embodiment of the invention, the nanofibers have an average diameter of less than 0.1 micron and the scaffold particles have a first dimension of greater than about 1 µm and a second dimension of greater than about 0.5 µm.

According to one embodiment of the invention, the packed bed further comprises an additive (e.g. a particulate additive) as a third component which is incorporated into the bed in an amount ranging from 0.01 wt% to 50 wt%, more preferably from 0.01 to 20 wt%, even more preferably from 0.01 to 10 wt%, and most preferably from 0.01 to 5 wt%.

Manufacturing process of nanofiber filling beds

Generally, the process involves mixing the nanofibers with the scaffold particles to form a packed bed structure. Suitable comparable processes are set forth in U.S. Patent Application Serial No. 08/428,496, filed April 27, 1995, hereby incorporated by reference.

According to one embodiment of the invention, a porous composite filling bed with a plurality of nanofibers and a number of scaffold particles is produced by a method which comprises the following steps:

(a) dispersing the nanofibers and the scaffold particles in a medium to form a suspension; and

(b) Separating the medium from the suspension to form the filler layer.

According to a preferred embodiment of the invention, packing beds were prepared using either glass fibers made from commercial glass fibers (e.g. Whatman® GF/B glass fiber filter paper) or chopped carbon fibers as scaffold particles. In both cases, the nanofibers (graphite fibrils) and the larger fibers were mixed together to form suspensions of the two and filtered through a vacuum filter port to remove the fluid and form a dry, packed "mat" or bed.

The medium is preferably selected from the group consisting of water and organic solvents.

The separation step may comprise filtering the suspension or evaporating the medium from the suspension.

According to one embodiment, the suspension is a gel or a paste comprising the nanofibers and scaffold particles in a fluid. The separation step may comprise the following steps:

(a) heating the gel or paste in a pressure vessel to a temperature above the critical temperature of the fluid;

(b) removing the supercritical fluid from the pressure vessel; and

(c) Removing the porous packed bed composite from the pressure vessel.

In another embodiment, a slurry of carbon nanotubes is prepared and chopped or ground segments of scaffold particles are added to a suspension containing the nanofibers and agitated to keep the materials in dispersion. The medium is then separated from the suspension, forming a packed bed with good fluid flow properties. Fiber support particles 0.318 cm (1/8 inch) long and even smaller work well.

Method for using nanotube packing beds

The fibril packed beds can be used for purposes where porous media are known to be useful. These include filtration, electrodes, adsorbents, chromatography media, etc. In addition, packed beds are a suitable bulk form of nanofibers and can therefore be used for any known application including in particular EMI shielding, polymer composites, active electrodes, etc.

For some applications, such as EMI shielding, filtration, and current collection, unmodified nanofiber packing beds can be used. For other applications, the nanofiber packing beds are a component of a more complex material, that is, they are part of a composite. Examples of such composites are polymer casting mats, chromatography media, electrodes for fuel cells and batteries, and ceramic composites, including bioceramic materials such as artificial bone. In some of these composites, such as casting mats and artificial bone, it is desirable that the non-nanofiber components fill or substantially fill the porosity of the packing bed. For others, such as electrodes and chromatographic agents, their usefulness depends on the composite preserving at least some porosity of the packing bed.

The rigid networks can also serve as the backbone in biomimetic systems for molecular recognition. Such systems have been described in US Patent No. 5,110,833 and International Patent Publication No. WO93/19844.

The packed bed products described above can be incorporated into a matrix. Accordingly, a non-nanofiber component is filtered through the bed and solidified to form a composite. Preferably, the matrix is an organic polymer (e.g., a thermosetting resin such as an epoxy, bismaleimide, polyamide, or polyester resin; a thermoplastic resin; a reaction molding resin; or an elastomer such as natural rubber; styrene-butadiene rubber or cis-1,4-polybutadiene); an inorganic polymer (e.g. a polymeric inorganic oxide such as glass), a metal (e.g. lead or copper) or a ceramic material (e.g. Portland cement).

Filling beds as electrodes

According to a preferred embodiment, the packed beds described above are used as porous electrode materials and have been shown to be good porous electrode materials. Accordingly, another aspect of the invention relates to materials for flow electrodes comprising the nanofiber packed beds with improved fluid flow properties. The porous electrode material is made from the packed bed mixtures of the nanofibers and scaffold particles. The electrode materials can advantageously utilize the composite properties of different materials in the composite mixture. In addition, the conductivity of the system can be increased when different conductive scaffold particles are mixed with the nanofibers.

According to one embodiment, electrodes can be made by attaching a conductive wire to areas of the packing beds with conductive epoxy. Such electrodes can be examined by cyclic voltammetry. It was observed that the material filtered better than the fibrils alone would have done.

According to a further embodiment, an electrical connection is formed by forming the packed bed on a conductive surface. For example, the packed bed can be formed by filtering the packed bed onto a conductive screen. Preferably, a gold, platinum, nickel or stainless steel screen or filter screen is used.

In addition, the conductivity can be limited by the internal connection of the nanofibers. Composites with larger fibers mixed into the nanofibers allow for improved porosity and/or flow properties as well as improved conductivity. This allows the surface of the nanofibers to be more easily utilized in electrode applications in situations where both conductivity and fluid flow through the material are required.

The addition of varying amounts of larger framework particles with a fixed amount of nanofibers allows the internal volume to be precisely tuned. The small pore sizes in the structure lead to an electrochemical thin film behavior, whereby the effective "thickness" of the film layer effect can be varied. Consequently, porous, highly conductive electrodes with different effective void volumes can be made.

EXAMPLES

The following examples illustrate some of the products and processes for making them which fall within the scope of the present invention. Of course, they are not to be considered as limiting the invention in any way. Numerous changes and modifications can be made to the invention.

Example I (Comparison) Production of the fibril mat using known methods

A diluted fibril dispersion is used to make the mats or layers. A fibril suspension containing 0.5% fibrils in water is prepared using a Waring blender. Once diluted to 0.1%, the fibrils are further dispersed using a sonication probe. The dispersion is then vacuum filtered to form a mat, which is then oven dried.

The mat has a thickness of about 0.20 mm and a density of about 0.20 gm/cc, corresponding to a pore volume of 0.90. The electrical resistance in the plane of the mat is about 0.02 ohm/cm. The resistance perpendicular to the mat is about 1.0 ohm/cm. The fluid flow properties of the mat are poor.

Example II (comparison) Production of the fibril mat using known methods

A fibril suspension containing 0.5% fibrils in ethanol is prepared using a Waring blender. After dilution to 0.1%, the fibrils are further dispersed using a sonication probe. The ethanol is then allowed to evaporate and a mat is formed. The fluid flow properties of the mat are poor.

Example III (comparison) Preparation of porous fibril plugs using known methods

The removal of a supercritical fluid from a well-dispersed fibril paste is used to produce low density molds. 50 cc of a 0.5% dispersion in n-pentane is added to a slightly larger capacity pressure vessel equipped with a needle valve to allow slow pressure release. After the vessel is heated to a temperature above the critical temperature of pentane (Tc = 196.6ºC), the needle valve is opened slightly to release the supercritical pentane over a period of about one hour.

The resulting solid fibril plug, which has the shape of the vessel interior, has a density of 0.005 g/cc, corresponding to a pore volume fraction of 0.997%.

The resistance is isotropic and is approximately 20 ohms/cm. The mat has poor fluid flow properties.

Example IV (according to the invention) Production of composite mats from fibrils and glass fibers

A fibril suspension was prepared by mixing 2 g of fibrils in 400 mL of deionized water (0.5% w/w) and mixing in a Waring® blender for 5 minutes on high power. 60 mL of the suspension was diluted to 200 mL with deionized water and sonicated with a 450 watt Branson® sonicator probe for 13 minutes on high power with 100% operation to prepare a fibril stock dispersion. Disks (0.25 inch diameter, 6.6 mg/disk) were punched out of a Whatman® GF/B glass fiber filter. Aliquots of the fibril stock dispersion and the glass fiber disks were mixed in the proportions indicated in Table 1: TABLE 1: Suspension containing carbon nanotubes and glass fiber particles

Each 100 ml mixture was sonicated for an additional 4 minutes at high intensity and passed through a 0.45 μm MSI nylon filter in a 47 mm Millipore membrane filter manifold The fibril/glass fiber composites formed felt-like mats which were peeled from the nylon membranes and dried for several hours at 80ºC between two pieces of filter paper under weight to maintain flatness. The area of each mat disk, defined by the dimensions of the filtration manifold, was 10 cm². Each mat was weighed and the thickness was measured with micrometer screws. The data are listed in Table 2: TABLE 2: Data on a mixed glass fiber/CC fibril mat

* 6.6 mg/slice

** set at a total of 36 mg/cc Glass d" and fibril d" refer to the total weight of the fibril or glass fiber divided by the total volume of the mat.

The flow characteristics of each mat were measured by observing the flow of water through each mat in a 25 mm diameter membrane filter manifold connected to a vacuum pump capable of producing a vacuum of approximately 0.982 x 10⁴ Pa (29"Hg). The 10 cm2 mats were centered and clamped in the 25 mm diameter filter manifold to ensure that fluid could only flow through the mat and not outside the perimeter. The deionized water used for the flow-through study was filtered sequentially through a 0.45 µm pore size nylon filter, a 0.2 µm pore size cellulose membrane filter, and finally a 0.1 µm pore size cellulose membrane filter to remove trace materials that might interfere with the flow-through studies. For each mat, the reservoir on the manifold was filled to above the 15 mL line and a vacuum was applied to ensure flow-through. When the meniscus of the upper water level passed the 15 mL mark, a timing device was started and the time taken for the fluid level to return was recorded until the 5 mL level was reached. The active filtration area was approximately 2 cm2. In this way, the liquid flow through was measured at several times during the passage of 10 ml of water. This procedure was used for each composite mat as well as for a layer of Whatman® GF/B glass fiber filter. The flow through data is listed in Table 3: TABLE 3: Measured times* versus volume passed through for fibril/glass fiber mats

"-" means not measured

* Time in minutes

For each filter mat, the throughput was linear versus time, indicating that the mats were not plugged and that the throughput rate was constant over time. It is also observed that the throughput rate increases with increasing proportions of glass fibers in the composite mats. As can be seen from the data in Table 3, the throughput rate observed for the composite mat No. 184-20-5 is almost two orders of magnitude greater than that of mat No. 184-20-1, a mat consisting only of fibrils.

Example V (according to the invention) Production of composite mats from fibrils and carbon fibres

Composite mats of fibrils and chopped carbon fibers were prepared and tested for flow properties. A fibril suspension was prepared by adding 2 g of fibrils to 400 mL of deionized water (0.5% w/w) and mixing in a Waring® blender for 5 minutes on high. 70 mL of the suspension was added to 280 mL of deionized water, a magnetic stir bar was added, and the suspension was sonicated with a 450 watt Bransonic® sonicator probe at constant agitation for 20 minutes to more completely disperse the fibrils and produce 350 mL of a 0.1% w/w stock dispersion. Weighed amounts of chopped carbon fibers with a cut length of 0.3175 cm (1/8 inch), Renoves, ex-PAN fibers, were placed in a series of beakers (seven) in the proportions shown in Table 4. TABLE 4

100 mL of deionized water and 50 mL of the 0.1% fibril stock dispersion were added to each beaker. Each of the seven solutions was sonicated for 5 minutes with stirring and filtered through a 0.145 µm nylon filter in a 47 mm Milupore membrane filter manifold. The mats were left on the membranes, placed between pieces of filter paper, and dried overnight at 80 ºC between the two pieces of filter paper under weight to maintain flatness. The weight and thickness of each 10 cm2 mat was measured and the data (after subtracting the nylon membrane weight of 93.1 mg and thickness of 101.6 µm (4.0 mils)) are shown in Table 5 along with a listing of the proportional composition of each mat. TABLE 5

GF = graphite-like fibrils

d' GF = weight of the fibrils divided by the total volume of the mat

d' CF = weight of carbon fibers divided by the total volume of the mat

The fibril/carbon fiber mats still on the nylon membranes were used for permeation studies as described in the previous examples; the permeation data are shown in Table 6. The permeation through the 0.45 µm nylon filter itself is very high. Consequently, its presence does not significantly alter the permeation measurements of the supported mats. TABLE 6

For each filter mat, the flow rate is linear versus time, indicating that the filters were not clogged and that the flow rate was constant over time. In comparison to flat fibril mats, it is observed that the flow rate shows a slight decrease with the two lower carbon fiber concentrations, but increases dramatically at higher carbon fiber fractions.

Example VI (according to the invention)

Use of Fibril/Glass Fiber Composite Mats as Electrodes Portions of the fibril/glass fiber composite mats used in Example IV, approximately 5 mm and 8 mm in size, were cut from the mats using a razor blade and formed into electrodes. Fabrication involved connecting a piece of copper wire to one end of each of the 5 mm by 8 mm sections with graphite paint (Ladd Industries) and insulating the contact point with epoxy. The copper wire was insulated with a glass sleeve and sealed with epoxy to the attached area of the mat, leaving only a 4 mm by 4 mm "panel" of composite mat exposed. Two electrodes were made from each mat as indicated in the table.

TABLE 7

Composite # Electrode #

184-20-1 184-21-1,2

184-20-2 184-21-3,4

184-20-3 184-21-5,6

184-20-4 184-21-7,8

184-20-5 184-21-9,10

The composite electrodes were examined by cyclic voltammetry using an EG&G PAR 273 potentiostat, an Ag/AgCl reference electrode (Bioanalytical Systems, Inc.), and a Pt mesh counter electrode in a single-chamber cell (Bioanalytical Systems, Inc.) filled with a solution containing 3 mM potassium ferrocyanide, 3 mM potassium ferrocyanide, and 0.5 M K2SO4 in water. The cyclic voltamograms and ferri-/ferrocyanide voltammetric curves, respectively, showed oxidation and reduction waves with properties that varied with the composition of the electrodes. Characteristic features of the cyclic voltamograms, recorded at a scan rate of 25 mv/second, are listed in Table 8. TABLE 8

* Thickness in mils

** Area in cm²

Electrode 184-21-1, which consisted only of fibrils, showed very sharp peaks with minimal peak separation, consistent with a redox process occurring in a porous electrode with very small pores. The peak shape and separation between the anodic and cathodic peaks was similar to that observed in a thin film cell. The dependence of the anodic peak current on scan speed shows a nearly linear dependence for electrode 184-21-1, but becomes increasingly nonlinear as the proportion of glass fibers in the electrode material increases. The anodic peak currents in milliamperes recorded at various scan speeds are shown in Table 9. TABLE 9: Peak anode current (IPA), mA vs. scan speed, mv/sec

Example VII (according to the invention) Use of fibril/carbon fiber composite mats as electrodes

Composite mats of the fibrils and carbon fibers were used as electrodes. Carbon fibers are electrically conductive. Using the procedure described above in the previous examples, three fibril/carbon fiber composite mats were prepared. The proportions are shown in Table 10. TABLE 10: Composition of the fibril/carbon fiber mats used for the electrodes

Fibril/carbon fiber mat electrodes were prepared using the procedure described in Example VI. Electrode dimensions are listed in Table 11 below. TABLE 11

The composite electrodes were examined by cyclic voltammetry using an EG&G PAR 273 potentiostat, an Ag/AgCl reference electrode (Bioanalytical Systems, Inc.), and a Pt mesh counter electrode in a single-chamber cell (Bioanalytical Systems, Inc.) filled with approximately 15 ml of a solution containing 3 mM potassium ferricyanide, 3 mM potassium ferrocyanide, and 0.5 M K2SO4 in water. The cyclic voltamograms of ferri-/ferrocyanide showed oxidation and reduction waves with properties that varied with the composition of the electrodes. Characteristic features of the cyclic voltamograms are listed in Table 12. TABLE 12: Summary of cyclic voltammetry data of ferri-/ferrocyanide for fibril/carbon fiber mat electrodes

EPA = anodic peak potential at 25 mv/second, V versus Ag/AgCl

EPC = cathodic peak potential at 25 mV/second, V versus Ag/AgCl

EP-P = peak separation at 25 mv/second, mv

IPA = anodic peak current

All three carbon nanotube/carbon fiber composite electrodes showed very sharp oxidation and reduction peaks with minimal peak separation, which is consistent with a redox process taking place in porous electrodes with very small pores. The peak shape and separation between anodic and cathodic peaks is similar to that observed for a thin film cell. The dependence of the anodic peak current on the scan speed shows a nearly linear dependence for electrode 184-28-1, but some deviation from the linear dependence is observed for electrodes 184-28-3 and 184-28-4. Each composite mat has the same amount of carbon nanotubes per unit area, but the addition of the carbon fibers to the composite leads to an increase in thickness and thus in the electrode volume. The peak currents and the integrated Currents increase with electrode volume, as expected due to the increased amounts of ferric/ferrocyanide solution in the porous electrode.

These results demonstrate that the porosity of fibril mat electrodes can be modified by forming composites with larger diameter fibers, allowing access to larger amounts of material in the solution phase. Furthermore, the use of conductive carbon fibers preserves the conductive nature of the fibril mats by diluting the fibrils in the mats with the larger carbon fibers in the composite.

The carbon fibers have a diameter of approximately 7-8 µm and the fibrils have a diameter of approximately 0.01 µm. Thus, although the fibril mat used to prepare electrode 184-28-4 contains 80% carbon fibers (w/w), the carbon fibers contribute little to the total surface area of the electrode. This is confirmed by measurements of the double layer capacitance of the charging currents, which correlate with the electrochemically accessible surface area. The double layer charging currents were measured by recording cyclic voltammograms at 10 mv/second in an electrolyte containing only 0.5 M KaSO₄ in water. Half of the total current difference between the cathodic and anodic sweeps of the cyclic voltammograms measured at 0.0 V versus Ag/AgCl was used as the double layer charging current (Ids). As can be seen from the data in Table 13, the double-layer charging current, normalized with respect to the electrode area and therefore with respect to the fibril mat, is almost constant, although the carbon fiber content of the electrode materials varies over a wide range. TABLE 13

Example VIII (according to the invention) Use of fibril mats as flow electrodes

A 50 mL suspension of carbon nanotubes in water was prepared at a concentration of 1 mg per milliliter. The suspension was sonicated using a 450 watt Branson sonicator probe at full power with 20% operation for 20 minutes to ensure that the carbon nanotubes were well dispersed. The dispersion was vacuum filtered through a 0.45 µm MSI nylon filter in a 47 mm Millipore membrane filter manifold. The carbon nanotubes formed a felt-like mat, which was peeled off the nylon membrane and dried between two pieces of filter paper at 80ºC for two hours using weight to maintain flatness. The thickness (or height) of the dried carbon nanotube mat was 8 mils as measured with micrometer screws. A 13 mm arc-shaped punch was used to produce a disk of the carbon nanotube mat with a diameter of 13 mm.

An electrochemical flow cell was constructed from a 13 mm Swinney-type plastic membrane filter holder by attaching a disk of 13 mm diameter gold mesh (400 mesh, Ladd Industries) to the top of the membrane support and making an electrical contact to the filter mesh with a platinum wire insulated with Teflon® shrink tubing that passed through the wall of the filter holder for external connection as the working electrode of a three-electrode potentiostat circuit. The gold screen was fixed in place with a minimal amount of epoxy resin around the outer edge. A strip of gold foil was formed into a ring and attached to the bottom, outlet side of the filter holder and connected to a platinum wire lead fed through the wall of the filter holder for external connections as the counter electrode of the three-electrode potentiostat circuit. A 0.5 mm diameter silver wire ring was electrochemically oxidized in 1M HCl, washed with water and attached to the top of the filter holder with the end of the wire fed out through the wall for external connections as the reference electrode in the three-electrode potentiostat circuit. The appropriate external contacts on the flow cell were connected to the working, counter and reference leads of an EG&G PAR 273 potentiostat. The flow cell was connected to a Sage syringe pump with interchangeable syringes for using different solutions. Initial background measurements were made in 0.5 M KiSO4 by examining the cyclic voltammetric response of the gold screen both at rest and when flowing at 0.6 mL/minute. A solution containing 2.5 mM potassium ferricyanide, 2.5 mM potassium ferrocyanide, 10 mM KCl, and 0.5 M K2SO4 in water was switched to and cyclic voltammograms were recorded under static and flow conditions at 0.4 mL/minute. There was a difference of 0.225 mA between the anodic and cathodic peak currents at a scan rate of 10 mV/second under static conditions with no flow. These control experiments demonstrated that the reference electrode was stable under flow conditions and established the proportions of the background current attributable to the gold screen support.

The 13 mm carbon nanotube mat disk was placed on top of the gold screen, followed by the gasket fitted with the filter holder and the upper portion of the filter holder. The cell was filled with a solution containing 2.5 mM potassium ferricyanide, 2.5 mM potassium ferrocyanide, 10 mM KCl and 0.5 M K₂SO₄ in water. Cyclic voltamograms for the reduction and oxidation of ferri-/ferrocyanide recorded under static conditions showed peak currents that were nearly 50 times larger than those observed for the gold mesh alone. There was a difference of 12 mA between the anodic and cathodic peak currents at a scan rate of 10 mv/second under static conditions without sweeping. In addition, the shape of the cyclic voltamograms recorded at scan rates of 10, 20, and 40 mv/second was consistent with the oxidation and reduction of ferri-/ferrocyanide trapped in the pore structure of the carbon nanotube mat electrode. Similar shapes for cyclic voltammetry were recorded under flow conditions at a pump rate of 0.4 ml/minute. The results are shown in Figure 8, which is a graphical representation of the current/voltage relationship for the materials prepared according to Example 8, where the vertical axis represents the current and the horizontal axis represents the applied potential at several potential scanning speeds.

The terms and expressions employed are used as terms of description in a non-limiting sense and the use of such terms and expressions is not intended to exclude equivalents of features shown and described; it will be recognized that various modifications are possible within the scope of the invention.

In summary, the invention comprises a composition comprising a porous packed bed comprising a plurality of nanofibers and a number of scaffold particles, the packed bed having a water fluid flow rate characteristic of greater than 0.5 ml/minute/cm2 at a pressure differential across the packed bed of about 1 atm when the packed bed has a thickness of 25.4 µm (one mil). The composition may comprise a packed bed having a surface area of greater than about 10 m2/g. The composition may Nanofibers having an average diameter of less than 0.1 µm and scaffold particles having a first dimension of greater than about 1 µm and a second dimension of greater than about 0.5 µm.

The composition may comprise a packed bed comprising 5 wt.% to 50 wt.% nanofibers and 90 wt.% to 60 wt.% scaffold particles. The composition may comprise a packed bed comprising a solid non-void volume containing 1 vol.% to 99 vol.% nanofibers and 99 vol.% to 1 vol.% scaffold particles. The composition may comprise a packed bed comprising a solid non-void volume containing 5 vol.% to 50 vol.% nanofibers and 90 vol.% to 60 vol.% scaffold particles.

The composition may comprise scaffold particles having a shape and a size suitable for causing a scaffold effect in the packing bed. The composition may comprise scaffold particles having a specific shape selected from a fiber, an irregular solid, a sphere, a plate, a disk, a pyramid or a cube. The composition may comprise nanofibers and scaffold particles, these comprising carbon.

The composition may further comprise particulate additives dispersed in the packed bed. The composition may comprise particulate additives present in an amount in the range of 0.01 to 25% by weight.

This invention also relates to a method for producing a porous filled bed composite with a plurality of nanofibers and a number of scaffold particles, which comprises the following steps:

(a) dispersing the nanofibers and the scaffold particles in a medium to form a suspension; and

(b) Separating the medium from the suspension to form the packed bed.

The method may comprise a medium selected from the group consisting of water and organic solvents.

This invention further relates to a porous flow electrode comprising a porous packed bed having a plurality of nanofibers and a number of scaffold particles, wherein the packed bed has a porosity of greater than 50% and a water flow rate characteristic of greater than 0.5 ml/minute/cm2/mil thickness at a pressure through the packed bed of about 1 atm when the packed bed has a thickness of 25.4 µm (1 mil).

The porous flow electrode may comprise a means for electrically connecting the electrode, which is a conductive wire or a conductive planar surface in conductive contact with the electrode. The porous flow electrode may comprise a means for electrically connecting the electrode, which is a current collection device that provides a support for the packed bed.

Claims (20)

1. A composition of matter comprising a porous packed bed containing a plurality of nanofibers and a number of scaffold particles, wherein the packed bed has a water fluid flow rate index of greater than 0.5 ml/min/cm2 at a pressure differential across the packed bed of about 1 atm when the packed bed has a thickness of 25.4 µm (1 mil). 2. The composition of claim 1, wherein the packed bed has at least one fluid flow index that is at least twice that of a nanofiber packed bed without scaffold particles. 3. A composition according to claim 1 or 2, wherein the packed bed comprises 1 wt.% to 99 wt.% nanofibers and 99 wt.% to 1 wt.% scaffold particles. 4. A composition according to any preceding claim, wherein the nanofibers have a diameter of less than about 0.5 µm and a length/diameter ratio greater than about 5. 5. A composition according to any preceding claim, wherein the nanofibers are carbon fibrils that are substantially cylindrical with a substantially constant diameter of less than 0.5 µm and have graphite layers that are concentric with respect to the fibril axis and substantially free of pyrolytically deposited carbon. 6. A composition according to any preceding claim, wherein the scaffold particles are fibers having an average diameter that is at least 10 times larger than the average diameter of the nanofibers. 7. A composition according to any preceding claim, wherein the nanofibers are carbon nanofibers and the particles are carbon fibers. 8. A composition according to any preceding claim, wherein the packed bed has substantially isotropic physical properties in at least two dimensions. 9. A method for producing a porous packed bed composite with a plurality of nanofibers and a number of scaffold particles, comprising the following steps: (a) dispersing the nanofibers and the scaffold particles in a medium to form a suspension; and (b) Separating the medium from the suspension to form the packed bed. 10. The method of claim 9, wherein the separating step comprises filtering the suspension. 11. A method according to claim 9 or 10, wherein the separating step comprises evaporating the medium from the suspension. 12. A method according to any one of claims 9, 10 or 11, in which the suspension is a gel or a paste comprising the nanofibers and scaffold particles in a fluid and the separation comprises the following steps: (a) heating the gel or paste in a pressure vessel to a temperature above the critical temperature of the fluid; (b) removing the supercritical fluid from the pressure vessel; and (c) Removing the porous packed bed composite from the pressure vessel. 13. A porous flow electrode comprising a porous packed bed containing a plurality of nanofibers and a number of scaffold particles, the packed bed having a porosity of greater than 50% and a liquid flow rate index for water of greater than 0.5 ml/min/cm2/25.4 µm thickness at a pressure of about 1 atm through the packed bed when the packed bed has a thickness of 25.4 µm (1 mil). 14. A porous flow electrode according to claim 13, further comprising a means for electrically connecting the electrode. 15. Porous flow electrode according to claim 13 or 14, in which the packed bed is supported on a conductive mesh. 16. A method for producing a porous flow electrode comprising a porous packed bed with a plurality of nanofibers and a number of scaffold particles, comprising the following steps: (c) dispersing the nanofibers and the scaffold particles in a medium to form a suspension; and (d) separating the medium from the suspension to form the packed bed, wherein the separating step comprises filtering the medium through a support comprising a conductive mesh to lead to the porous electrode which is in conductive contact with the conductive mesh. 17. A filter medium comprising a composition according to any one of claims 1 to 8. 18. Adsorption medium comprising a composition according to any one of claims 1 to 8. 19. Solid phase for chromatographic separation comprising a composition according to any one of claims 1 to 8. 20. A composition comprising a porous packed bed having a plurality of nanofibers and a number of scaffold particles, the packed bed having a liquid flow rate index for water such that F > ((0.5 ml · 25.4 um/min · atm · cm²) · ((P · A) / T)), where F is the flow rate through the packed bed in mL/min, P is the pressure difference across the packed bed in atmospheres, A is the area of the packed bed in cm² and T is the thickness of the bed in µm.